Motion of an object can be tracked based on video streams from a stationary camera. However, accurate tracking of motion generally requires fairly high-resolution images as well as sophisticated and computationally intense algorithms. When images are not recorded at a significantly high resolution and/or computational resources are limited, this can significantly reduce the ability to accurately track motion.